Biological Conservation 98 (2001) 241±249 www.elsevier.com/locate/biocon

Evidence for spider community resilience to invasion by non-native spiders

Jutta C. Burger a,*, Michael A. Patten b, Thomas R. Prentice a, Richard A. Redak c aDepartment of Entomology, University of California, Riverside, CA 92521, USA bDepartment of Biology, University of California, Riverside, CA 92521, USA cDepartment of Entomology and Center for Conservation Biology, University of California, Riverside, CA 92521, USA

Received 14 June 2000; received in revised form 20 June 2000; accepted 17 August 2000

Abstract The negative impacts of non-native species are well documented; however, the ecological outcomes of invasions can vary widely. In order to determine the resilience of local communities to invasion by non-native spiders, we compared spider assemblages from areas with varying numbers of non-native spiders in California coastal sage scrub. Spiders were collected from pitfall traps over 2 years. Productive lowland coastal sites contained both the highest proportion of non-natives and the greatest number of spiders overall. We detected no negative associations between native and non-native spiders and therefore suggest that non-native spiders are not presently impacting local ground-dwelling spiders. Strong positive correlations between abundances of some natives and non-natives may be the result of similar habitat preferences or of facilitation between species. We propose that the e€ects of non- native species depend on resource availability and site productivity, which, in turn, a€ect community resilience. Our results support the contention that both invasibility and resilience are higher in diverse, highly linked communities with high resource availability rather than the classical view that species poor communities are more invasible. # 2001 Elsevier Science Ltd. All rights reserved.

Keywords: Non-native species; Biological invasions; Invasibility; Community resilience; Arthropoda; Araneae

1. Introduction on population levels (but may or may not alter foraging behavior, habitat choice, etc.). That an entire ecosystem Non-native species frequently pose a threat to the can be a€ected, directly or indirectly, by an exotic spe- persistence of natives, presumably through competitive cies is exempli®ed by such classic cases as the brown tree exclusion or depredation. Several examples of the snake Boiga irregularis in Guam (Fritts and Rodda, devastating, community-wide e€ects of introduced 1998), the red imported ®re ant Solenopsis invicta in organisms exist (e.g., Enserink, 1999, Strayer, 1999), but Texas (Porter and Savignano, 1990), and cheatgrass these examples represent extreme cases. More com- Bromus tectorum in shrub-steppe vegetation (Brandt monly, we encounter non-native species seemingly and Rickard, 1994). The e€ects that non-native species coexisting with natives, and for the most part we are can have on populations of unrelated species have also unaware of their e€ects (Aniansson, 1999). been documented, as with European Starlings Sturnus Non-native organisms potentially exert four distinct vulgaris outcompeting nesting Northern Flickers Cola- types of e€ects on a native species or community. They petes auratus (Ingold, 1996) and Purple Martins Progne may (1) impact and degrade an entire community or subis (Brown, 1997). Reports of negative e€ects on ecosystem; (2) cause declines in speci®c, possibly unre- populations of related species are numerous. The non- lated, native species or small sets of unrelated species native cray®sh Orconectes rusticus outcompetes native (presumably through competition or predation); (3) Orconectes species in Wisconsin (Hill and Lodge, 1999), cause declines in closely related taxa (most likely bullfrogs Rana catesbiana deplete populations of native through competition); or (4) have no discernible e€ect Rana species throughout California (Hayes and Jen- nings, 1986), and Argentine ants, Linepithema humile, * Corresponding author. depress many native ant species' populations along E-mail address: [email protected] (J.C. Burger). urban interfaces in California (Holway, 1999, Suarez et

0006-3207/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved. PII: S0006-3207(00)00159-2 242 J.C. Burger et al. / Biological Conservation 98 (2001) 241±249 al., 1998). Published cases of exotics not a€ecting natives that have become established in southern Cali- natives are less common (Bauchau, 1997), but do exist. fornia within the last 100 years (Chamberlin and Ivie, For example, non-native ®shes in some streams of 1935, Gertsch, 1979, Platnick and Shadab, 1983, Plat- northern California are not signi®cantly changing native nick and Murphy, 1984). Four of these seven are gna- species assemblages (Baltz and Moyle, 1993), in contrast phosids, the most abundant and diverse group of native to their e€ects in other regions of North America ground-dwelling spiders in CSS (Prentice et al., 1998). (Moyle, 1995, Chapleau et al., 1997). They represent a single guild (sensu Root, 1967) of It is thus dicult to predict the degree to which non- wandering, generalist predators. One, Dysdera crocata natives are able to invade ecosystems. Elton (1958), (Dysderidae), is an isopod specialist (Gertsch, 1979). MacArthur (1970), Stachowicz et al. (1999) and others Another, Oecobius annulipes (Oecobiidae), is a small, proposed that species-poor ecosystems are most easily relatively sessile web-builder. Yet another, Metaltella invaded. Such systems contain species depauperate simoni (Amphinectidae), is a moderately large ground communities that may re¯ect low resource availability dweller. or high levels of disturbance and result in a less ecient Based on generally accepted theory, if the hypothesis utilization of resource space. However, the bulk of of higher invasibility of low diversity sites is true, then experimental data suggest the opposite trend: areas of sites with lower species richness should have more non- high diversity are more likely to support invading spe- natives (and, by corollary, vice versa). Thus, our objec- cies (Levine and D'Antonio, 1999, Levin, 2000). Diver- tives were (1) to describe the pattern of occurrence of sity may re¯ect high resource availability, high non-native spiders across sites in CSS, (2) to determine productivity, and high structural complexity. the degree of association between native spiders, non- Similar disagreement exists as to which conditions native spiders, and their environment, and (3) to study most favor large scale e€ects on resident fauna (Levine the intraguild relationships between closely related and D'Antonio, 1999). Whatever the optimal conditions native and exotic gnaphosid spiders. We predicted that, for invasion and establishment may be, it is certain that if non-natives were impacting native spider commu- habitat fragmentation and anthropogenic in¯uences nities, their e€ects would be most obvious within the increase both the likelihood of invasion and the severity ecologically and behaviorally similar gnaphosids. of the invaders' e€ects on natives (Enserink, 1999). We have studied geographic patterns in local spider assem- blages in order to determine which of these processes 2. Methods may apply to ground-dwelling spider communities in southern Californian coastal sage scrub (CSS) habitats. 2.1. Field collection and identi®cation CSS is a major vegetation type in cismontane south- ern California (Westman, 1981). Its canopy is composed In order to describe patterns of abundances of native of drought-deciduous shrubs such as Artemisia cali- and non-native spiders, we sampled spiders from 60 fornica (Asteraceae) and Eriogonum fasciculatum (Poly- sites in undisturbed CSS and chaparral intergrades gonaceae). Its understory was historically composed of across coastal San Diego County, California, USA. a rich array of native annual forbs. Primarily as a result Sites were located at Marine Corps Base Camp Pendle- of urban development, associated changes in ®re fre- ton and Marine Corps Air Station Miramar (Prentice et quency, and increased pollution, CSS has become al., 1998). Spiders were collected during 1-month sam- increasingly fragmented and vulnerable to invasion by pling periods in December 1994 and 1995, May 1995 non-native plant and animal species (Westman, 1979, and 1996, and August 1995 from ®ve randomly placed Zedler et al., 1983, Stylinski and Allen, 1999). Both 473 ml (16 oz.) pitfall traps at each site. Traps remained non-native grasses (e.g. Bromus spp., Avena spp.) in place for the duration of the study. Collection cups and annual forbs (e.g. Erodium spp., Centaurea sol- contained a dilute aqueous solution of AlconoxTM stitialis) are now common in most CSS and are detergent and salt to increase miscibility and decrease threatening native species' persistence (Eliason and decay. Pitfall traps have been shown to be e€ective Allen, 1997). To our knowledge, only one invasive when estimating cursorial spider (and other arthropod) arthropod species, the Argentine ant, has been studied fauna richness and activity (Uetz and Unzicker, 1976) in CSS. This species outcompetes and eliminates native and are currently the most accepted method for con- ants in CSS within valleys along urban edges (Suarez ducting ecological studies on these species. Traps were et al., 1998). left open for 7 days during each collection period. All CSS contains a diverse spider fauna whose species mature (and distinct immature) spiders collected were composition and response to non-native invasion is lar- identi®ed to species by Prentice. All immature mor- gely unknown. We have documented over 200 spider phospecies that could not be identi®ed to a distinct species from CSS vegetation in San Diego County (Pre- genus or species with certainty were eliminated from the ntice et al., 1998). Of these species, seven are non- data set (17% of all morphospecies; 15% of all spiders). J.C. Burger et al. / Biological Conservation 98 (2001) 241±249 243

Taxonomy and nomenclature of Araneae follows Pren- We used detrended correspondence analysis (DCA; tice et al. (1998). Hill and Gauch, 1980; using PC-ORD) to investigate A single adult of the non-native gnaphosid Trachyze- whether the composition of native spider communities lotes barbatus (®rst San Diego County record, see Pren- (presence/absence of species) changed across an envir- tice et al., 1998) and of Metaltella simoni were collected. onmental gradient and whether that gradient corre- These records were eliminated from Mantel and detren- sponded to the level of invasion by non-natives. We ded correspondence analyses (see below), because they identi®ed three levels of invasion based on natural could not re¯ect patterns of occurrence. Several imma- breakpoints in non-native abundance at sites: `low' with ture Trachyzelotes that could not be identi®ed to species 0±14 non-native individuals, `med' with 15±59 indivi- also were collected. As no native Trachyzelotes are duals, and `high' with 60±148 individuals. DCA axis known from North America (Platnick and Murphy, scores were correlated (using Pearson product-moment 1984) and we collected only one other Trachyzelotes correlations) with environmental parameters (as de®ned species in our study, all immature Trachyzelotes were above). Those parameters most highly correlated with considered to be the more commonly collected Trachy- axis scores were used to identify axes 1 and 2. zelotes lyonetti. In total, we analyzed the patterns of We used logistic regression (Sokal and Rohlf, 1995; occurrence of 145 spider species, including the ®ve non- SAS Statistical Software, ver. 6.13, 1997) to account for native species for which we had sucient sample. All environmental e€ects on variation in the proportion of data were pooled across years and seasons in order to natives before testing for the association between examine overall e€ects on species numbers and to mini- natives and non-natives that was independent of envir- mize the number of zero values in any particular sam- onment. We transformed (arcsine square-root) the pling period. environmental variables represented by grass, forb, lit- We measured vegetation and geographic variables at ter, and shrub cover frequencies to normalize their dis- all sites to describe and account for the pattern of native tributions. Other variables (distance to coast, elevation, and non-native spider species occurrence explained by and aspect) were not transformed. Our baseline model local habitat conditions. Transects were established at used all environmental variables as a predictor of native ®xed locations within each plot and vegetation was species proportion. Once constructed, we used log-like- recorded in April±May 1995 from a total of 95 points lihood ratio tests to investigate the degree to which each placed randomly at 2-m intervals along transects using a non-native species increased the predictive power of the point-transect method. From these measurements we model built from environmental variables. All environ- obtained cover estimates for forbs, grasses, leaf litter, mental variables were retained in the model in order to and shrubs. We also measured distance to coast, aspect, maximize the variance explained in the proportion of and elevation using standard GIS software (ArcView, natives at sites. If an exotic species did not add to the ver. 3.2, 1999, ESRI) and electronic maps provided by power of the model predicting the proportion of natives, Marine Corps Base Camp Pendleton and Marine Corps we would conclude that that species had no signi®cant Air Station Miramar. Collectively, we refer to the vege- detectable e€ect on natives. tation and geographic measurements as environmental Finally, we used Pearson product±moment correla- parameters. tions to look for associations between abundances of native and non-native gnaphosids (omitting species that 2.2. Analyses occurred in fewer than four sites). Signi®cance levels were adjusted using a Dunn-SÏ ida k correction. We used the presence or absence of non-natives at sites in f correlations (Sokal and Rohlf, 1995) to deter- mine whether non-native spiders tended to co-occur. 3. Results Signi®cance levels were adjusted using a Dunn-SÏ ida k [a0=1(1a)1/k] correction for k multiple comparisons 3.1. General patterns for non-native spiders (Sokal and Rohlf, 1995). We tested for similarity in pattern of occurrence Non-native spiders occurred on 51 of the 60 sites between native and non-native species groups using a sampled. At eight sites they constituted more than 50% Mantel test (Douglas and Endler, 1982; PC-ORD, ver. of all spiders collected (Fig. 1). Occurrences of the non- 3.0, 1997, MJM Software) that compared Jaccard simi- natives Urozelotes rusticus and Dysdera crocata were larity matrices of species presence/absence. We expected highly correlated with one another (f=0.41, P<0.01, that native species assemblages would be negatively a0=0.01). Both species tended to occur in more coastal, associated with non-natives. Values in the non-native mesic sites. Dysdera crocata occurrence was also corre- presence/absence matrix were transformed by adding 1 lated with that of Oecobius annulipes (f=0.14, P< to avoid a divide-by-zero error for sites that contained 0.01, a0=0.01); both species tended to occur in lowland no non-natives. and valley sites. Oecobius annulipes, a cosmopolitan 244 J.C. Burger et al. / Biological Conservation 98 (2001) 241±249

Fig. 1. Proportion of non-native spiders collected from pitfall traps. Spiders were collected at sites within (A) Marine Corps Base Camp Pendleton, and (B) Marine Corps Air Station Miramar, San Diego County, California (n=60 sites). species that has colonized undisturbed habitats across expectations, we found a signi®cant positive association the Southwest (Shear, 1970, Prentice et al., 1998), was between native and non-native species groups occurring the most common spider collected from pitfalls (com- on sites (Mantel standardized r=0.29, P=0.001). When prising 16% of all spiders). It was most abundant at low native-only species assemblages on sites were compared elevation coastal sites. by DCA, they were ordered along a gradient associated with litter cover as well as with distance to coast and 3.2. Association between natives, non-natives, and associated changes in elevation (Fig. 3). Sites containing environment high numbers of non-natives clustered together and occurred at the low end of the gradient for litter. How- Although native species richness was uncorrelated ever, those containing medium or low/no numbers of with non-native abundance (Fig. 2), associations non-natives remained unordered with respect to either between the two groups emerged. Contrary to our litter cover or distance to coast. J.C. Burger et al. / Biological Conservation 98 (2001) 241±249 245

Fig. 2. Native species richness as a function of non-native abundance in coastal sage scrub (CSS) at Camp Pendleton and Miramar, San Diego County, California (n=60 sites).

Fig. 4. (A) Total spider abundance and (B) relative abundance of non- natives in castal sage scrub (CSS) at Pendleton and Miramar, San Diego County, California, as related to distance from coast (n=60 sites).

occurrence of natives. Each non-native species was Fig. 3. Detrended correspondence analysis (DCA) of native spider assemblages in coastal sage scrub (CSS) at Camp Pendleton and Mir- added in turn to the above logistic model and tested for 2 amar, San Diego County, California. Sites are labeled by non-native improvement in the model R . By so doing, we tested spider abundance, with L=0±14 non-natives, M=15±74, H=75±148 for associations between species that were independent (axis 1 gradient length=2.8, n=60 sites). Axis labels were determined of their association with a particular habitat. Only the by the highest correlation(s) between DCA scores for axes I and II and abundance patterns of Dysdera crocata and Oecobius environmental variables. annulipes signi®cantly increased the predictive power of the logistic model (Table 1). After we accounted for the Using a logistic regression model, the proportion of e€ect of the environmental variables, based on their native species occurring on sites could be predicted with parameter estimates, both species were positively asso- our chosen environmental parameters (w2=51.56, ciated with the proportion of natives. d.f.=7, P<0.001). Distance to coast was the only parameter that contributed signi®cantly to the model 3.3. Intraguild associations between native and exotic (Wald's w2=17.76, d.f.=1, P<0.001, a0=0.001); how- gnaphosids ever, it was strongly correlated with elevation. Distance to coast also was signi®cantly negatively correlated with The Gnaphosidae were the most diverse family of both overall spider abundance and the proportion of spiders collected, represented by 22 species, of which non-native spiders (Fig. 4a,b). four were non-native. Because of their diversity and Once we accounted for the e€ects of environmental their similarity in foraging behavior, we predicted that, variables on the presence of native spiders, we assessed if we were to ®nd evidence of negative associations whether exotic spider species were associated with the between natives and exotics, we would ®nd them within 246 J.C. Burger et al. / Biological Conservation 98 (2001) 241±249

Table 1 association was positive, in that non-natives occurred Chi-squared test of improvement in a logistic regression model for where native spiders were both most abundant and most proportion of native spiders at sites using seven environmental para- diverse. meters with non-native species occurrence included (n=60 sites; k=7 environmental variables) 4.2. Potential causes of observed patterns Model 2 log L R2 w2 P

Environment 291.477 0.58 ± ± Although we have no record of the spider fauna of Environment+Trachyzelotes spp. 290.782 0.58 0.695 n.s.a CSS before the introduction of non-natives, we Environment+Urozelotes rusticus 289.642 0.60 1.835 n.s. observed only positive associations between native and Environment+Zelotes nilicola 290.366 0.58 1.111 n.s. non-native spiders. Large-scale negative changes to a Environment+Dysdera crocata 286.252 0.61 5.225 <0.05 native arthropod fauna after the introduction of a non- Environment+Oecobius annulipes 250.826 0.79 40.651 <0.001 native predator can and have occurred. For example, a n.s., not signi®cant the non-native predator Vespula vulgaris (Hymenoptera: Vespidae) appears to be reducing densities of native the gnaphosids. Seven correlations between gnaphosid spiders and insects (Harris and Oliver 1993), and in speciess were signi®cant; all were positive and three manipulative studies it has been found to signi®cantly involved non-natives (Table 2). The native species reduce numbers of orb weaving spiders (Araneae: Ara- Drassyllus lamprus and Micaria utahna were highly neidae) (Toft and Rees, 1998). positively correlated with non-native members of the The absence of a negative association between natives Gnaphosidae. and non-natives in our study may be the result of sev- eral factors. Spider prey availability may be so high in the coastal region of San Diego County that competi- 4. Discussion tion for food between and predation from other spiders are relatively low. Examples of reduced competition due 4.1. Documented interactions between non-native spiders to high resource availability exist. Non-native pre- and natives dacious carabid beetles colonize sites quickly but do not reduce native congener density (NiemalaÈ et al. 1997). Non-native spiders appear to a€ect natives more Similarly, the introduced Coccinella septempunctata strongly in simpler, less productive urban landscapes than (Coleoptera: Coccinellidae) severely lowers survival of in natural ecosystems. In the eastern United States, the the related native Coleomegilla maculata at low prey European theridiid Steatoda bipunctata appears to be dis- density, but has no e€ect on the native species at high placing the native Steatoda borealis along urban interfaces prey density (Obrycki et al. 1998). (Ny€eler et al., 1986). In urban areas of southern Cali- Spider activity and/or abundance typically increase in fornia, the introduced marbled cellar spider (Pholcidae: relation to arthropod prey availability (Nentwig, 1982). Holocnemus pluchei) may be displacing the native black Many spiders, being opportunistic hunters, utilize other widow (Theridiidae: Latrodectus hesperus; W. R. Ice- spiders as prey (Polis et al., 1989, Wise, 1993). Adding nogle personal communication). In this study, non- species to a resource-rich environment could theoreti- native spiders were abundant in CSS and, at some sites, cally facilitate higher overall abundance and richness of were numerically dominant, particularly in lowland, species (sensu Hector et al., 1999). The positive associa- coastal sites (Fig. 4). Despite their prevalence, we failed to tion that abundance of both Dysdera crocata and ®nd a pattern that suggested non-native spiders were Oecobius annulipes had with the proportion of natives depressing native populations. There was a relationship independent of their association with environment between exotic and native spiders, but interestingly this (Table 1) may re¯ect such facilitation.

Table 2 Pearson product-moment correlations between gnaphosid spider speciesa

Drassyllus saphes Drassyllus lamprus Micaria icenoglei Micaria utahna Nodocion eclecticus

Drassyllus fractus 0.43 n.s.b n.s. n.s. n.s. Haplodrassus signifer n.s. 0.44 n.s. n.s. n.s. Sergiolus angustus n.s. n.s. n.s. n.s. 0.43 Urozelotes rusticus n.s. n.s. n.s. 0.61 n.s. Zelotes nilicola n.s. 0.66 n.s. 0.55 n.s

a All correlation coecients presented were signi®cant at a0=0.001. The species Urozelotes rusticus and Zelotes nilicola are non-natives. b n.s., not signi®cant. J.C. Burger et al. / Biological Conservation 98 (2001) 241±249 247

Spider assemblages may also have already changed in tioning and allotopy has been documented for the wolf response to non-natives. We have no information as to spiders Pardosa ramulosa and Pardosa tuoba (Lycosi- the exact time of local invasion, the character of the cur- dae; Greenstone, 1980). In our study, mesic coastal sites sorial spider community before invasion, or the current may currently contain ample prey and substrate com- stablility of assemblages. However, we have been unable plexity such that potential negative interactions between to detect any evidence of competitive exclusion or local generalist non-native and native predators are mini- extirpation across a gradient of non-native abundance. mized. Whereas biological invasions have proved Our results support recent contentions suggesting that devastating in many systems (Enserink, 1999, Kaiser, highly diverse communities are more likely than simple 1999, Malako€, 1999) and indeed have seriously a€ec- systems to support invading species (Levine and D'Anto- ted southern Californian ecosystems (Eliason and Allen, nio, 1999, Levine, 2000). Colonization by non-natives 1997), we failed to detect a deleterious impact of non- does not necessarily mean that native species, commu- native spider species upon their native counterparts. nities, or ecosystems are adversely a€ected. Productivity Although we lack data from a manipulative experiment, can increase with diversity when species act as facil- regardless of mechanisms involved, our data support itators for one another rather than competitors within a the hypothesis that non-native species are more com- hypothetical ®nite niche space (Hector et al., 1999). mon in areas with high biodiversity (Levine and Highly linked, diverse natural systems may actually be D'Antonio, 1999), and thus do not support the hypoth- able to support the establishment of a few invading esis that non-natives will more readily establish in areas species while still withstanding their potential negative with low biodiversity. Highly linked, diverse commu- e€ects. For example, grass tussocks supporting a high nities may be more resilient to the deleterious e€ects of diversity of native plant species contain more non- non-natives (Levine, 2000). We see a need for the natives than those with low diversity (Levine, 2000). development of better predictive models (based on life history and ecology) of the impact, and control, of ever- 4.3. Associations within the Gnaphosidae more-common biological introductions.

Because the Gnaphosidae are numerically dominant in ground-dwelling spider assemblages of CSS and con- Acknowledgements tain several non-native species (which are behaviorally and ecologically similar to natives), we predicted native The Department of Defense Legacy Resources Man- members of the family would be more a€ected by pre- agement Program funded this research under Coopera- dation and/or competition from exotic gnaphosids than tive Agreement No. N68711-94-LT-4042 from would other spider taxa. However, the only signi®cant Southwest Division, Naval Facilities Engineering Com- relationships we detected between native and non-native mand to R.A.R. for arthropod surveys at Marine Corps gnaphosid species were positive (Table 2). In another Air Station Miramar and Marine Corps Base Camp study, we observed that the native Drassyllus insularis, Pendleton. Map topology was provided to us by Marine despite being otherwise common along the coast, was Corps Air Station Miramar and Marine Corps Base absent at three coastal sites supporting the highest Camp Pendleton and is the property of the United numbers of the non-native Urozelotes rusticus (Prentice States Marine Corps. C. Matthies, M. Misenhelter, E. et al., 1998), suggesting that competitive displacement Porter, T. Hosch, K. Campbell assisted in the lab and of natives may occur in this system at very high densities ®eld, and E. Bonilla, S. Kensinger, T. Gaines, L. of exotic gnaphosids. Nonetheless, at a regional scale Truong, and M. Bungton assisted in the lab. W. Ice- (coastal southern California, USA) the pattern we nogle contributed valuable natural history observations found of association between native and exotic gna- and insight from years of ®eld collection in southern phosids does not suggest that these exotic species cur- California. C. Dunning, R.G. Wright, and two anon- rently are negatively a€ecting their native congeners or ymous reviewers provided helpful comments on an ear- confamilials. lier version of this manuscript.

5. Conclusion References

Intraguild predation is common in spiders (Hallander, Aniansson, B.H., 1999. Biological invasion or increased diversity. 1970, Nentwig, 1986, Polis et al. 1989). Additionally, Flora och Flora 94, 49±61. competition can be strong for food resources, nest and Baltz, D.M., Moyle, P.B., 1993. Invasion resistance to introduced retreat sites, and hunting territory (Gertsch and Reich- species by a native assemblage of California stream ®shes. Ecologi- cal Applications 3, 246±255. ert, 1976, Schaefer, 1978, Reichert, 1981). Coexistence Bauchau, V., 1997. Is there a ®le drawer problem in biological of closely related species by means of resource parti- research? Oikos 79, 407±409. 248 J.C. Burger et al. / Biological Conservation 98 (2001) 241±249

Brandt, C.A., Rickard, W.H., 1994. Alien taxa in the North American Levine, J.M., 2000. Species diversity and biological invasions: relating shrub steppe four decades after cessation of livestock grazing and local process to community pattern. Science 288, 852±854. cultivation. Biological Conservation 68, 95±105. Levine, J.M., D'Antonio, C.M., 1999. Elton revisited: a review of evi- Brown, C.R., 1997. Purple Martin Progne subis. In: Poole, A.F., Gill, dence linking diversity and invasibility. Oikos 87, 15±26. F.B. (Eds.), Birds of North America, No. 287. Academy of Natural MacArthur, R.H., 1970. Species packing and competitive equilibrium Sciences, Philadelphia, and American Ornithologists' Union, for many species. Theoretical Population Biology 1, 1±11. Washington, D.C. Malako€, D., 1999. Fighting ®re with ®re. Science 285, 1841±1843. Chamberlin, R.V., Ivie, W., 1935. Nearctic spiders of the family Moyle, P.B., 1995. Conservation of native freshwater ®shes in the Urocteidae. Annals of the Entomological Society of America 28, Mediterranean-type climate of California, USA: a review. Biological 265±279. Conservation 72, 271±279. Chapleau, F., Flindlay, C.S., Szenasy, E., 1997. Impact of piscivorous Nentwig, W., 1982. Epigeic spiders, their potential prey and competi- ®sh introductions on ®sh species richness of small lakes in Gatineau tors: relationship between size and frequency. Oecologia 55, 130± Park, Quebec. Ecoscience 4, 259±268. 136. Douglas, M.E., Endler, J.A., 1982. Quantitative matrix comparisons Nentwig, W., 1986. Non-webbuilding spiders: prey specialists or gen- in ecological and evolutionary investigations. Journal of Theoretical eralists? Oecologia 69, 571±576. Biology 99, 777±795. NiemelaÈ , J., Spence, J.R., Carcamo, H., 1997. Establishment and Eliason, S.A., Allen, E.B., 1997. Exotic grass competition in suppres- interactions of carabid populations: an experiment with native and sing native shrubland re-establishment. Restoration Ecology 5, 245± introduced species. Ecography 20, 643±652. 255. Ny€eler, M., Dondale, C.D., Redner, J.H., 1986. Evidence for dis- Elton, C.S., 1958. The Ecology of Invasions by Animals and Plants. placement of a North American spider, Steatoda borealis (Hentz), Matheun, London. by the European species S. bipunctata (Linnaeus) (Araneae: Ther- Enserink, M., 1999. Biological invaders sweep in. Science 285, 1834± idiidae). Canadian Journal of Zoology 64, 867±874. 1836. Obrycki, J.J., Giles, K.L., Ormord, A.M., 1998. Interactions between Fritts, T.H., Rodda, G.H., 1998. The role of introduced species in the an introduced and indigenous coccinellid species at di€erent prey degradation of island ecosystems: A case history of Guam. Annual densities. Oecologia 117, 279±285. Review of Ecology and Systematics 29, 113±140. Platnick, N.I., Murphy, J.A., 1984. A revision of the spider genera Gertsch, W.J., 1979. American Spiders, 2nd Edition. Van Nostrand Trachyzelotes and Urozelotes (Araneae, Gnaphosidae). American Reinhold, New York. Museum Novitates, 2792. Gertsch, W.J., Riechert, S.E., 1976. The spatial and temporal parti- Platnick, N.I., Shadab, M.U., 1983. A revision of the American spi- tioning of a desert spider community, with descriptions of new spe- ders of the genus Zelotes (Araneae, Gnaphosidae). Bulletin of the cies. American Museum Novitiates 2604. American Museum of Natural History 174, 97±192. Greenstone, M.H., 1980. Contiguous allotopy of Pardosa ramulosa Polis, G.A., Myers, C.A., Holt, R.D., 1989. The ecology and evolution and Pardosa tuoba (Araneae: Lycosidae) in the San Francisco Bay of intraguild predation: potential competitors that eat each other. Region, and its implications for patterns of resource partitioning in Annual Review of Ecology and Systematics 21, 297±330. the genus. American Midland Naturalist 104, 305±311. Porter, S.D., Savignano, D.A., 1990. Invasion of polygyne ®re ants Hallander, H., 1970. Prey, cannibalism and microhabitat selection in decimates native ants and disrupts arthropod community. Ecology the wolf spiders Pardosa chelata O.F. Mueller and P. pullata Clerck. 71, 2095±2106. Oikos 21, 337±340. Prentice, T.R., Burger, J.C., Icenogle, W.R., Redak, R.A., 1998. Spi- Harris, R.J., Oliver, E.H., 1993. Prey diets and population densities of ders from Diegan coastal sage scrub (Arachnida: Araneae). Pan- wasps Vespula vulgaris and V. germanica in scrubland-pasture. New Paci®c Entomologist 74, 181±202. Zealand Journal of Ecology 17, 5±12. Riechert, S.E., 1981. The consequences of being territorial: spiders, a Hayes, M.P., Jennings, M.R., 1986. Decline of ranid frog species in case study. American Naturalist 117, 871±892. western North America: are Bullfrogs Rana catesbeiana responsible? Root, R.B., 1967. The niche exploitation pattern of the blue-gray Journal of Herpetology 20, 490±509. gnatcatcher. Ecological Monographs 37, 315±350. Hector, A., Schmid, B., Beierkuhnlein, C., Caldeira, M.C., Diemer, Schaefer, M., 1978. Some experiments on the regulation of population M., Dimitrakopoulos, P.G., Finn, J.A., Freitas, H., Giller, P.S., density in the spider Floronia bucculenta (Araneae: Linyphiidae). Good, J., Harris, R., HoÈ gberg, P., Huss-Danell, K., Joshi, J., Symposia of the Zoological Society of London 42, 203±210. Jumpponen, A., KoÈ rner, C., Leadley, P.W., Loreau, M., Minns, A., Shear, W.A., 1970. The spider family Oecobiidae in North America, Mulder, C.P.H., O'Donovan, G., Otway, S.J., Pereira, J.S., Prinz, Mexico, and the West Indies. Bulletin of the Museum of Compara- A., Read, D.J., Scherer-Lorenzen, M., Schulze, E-D., Sia- tive Zoology 140, 129±164. mantziouras, A-S.D., Spehn, E.M., Terry, A.C., Troumbis, A.Y., Sokal, R.R., Rohlf, F.J., 1995. Biometry, 3rd Edition. W. H. Freeman Woodward, F.I., Yachi, S., Lawton, J.H., 1999. Plant diversity and and Company, New York. productivity experiments in European grasslands. Science 286, Stachowicz, J.J., Whitlatch, R.B., Osman, R.W., 1999. Species diver- 1123±1127. sity and invasion resistance in a marine ecosystem. Science 286, Hill, M.O., Gauch, H.G., 1980. Detrended correspondence analysis: 1577±1579. an improved ordination technique. Vegetatio 42, 47±58. Strayer, D.L., 1999. E€ects of alien species on freshwater molusks in Hill, A.M., Lodge, D.M., 1999. Replacement of resident cray®shes by North America. Journal of the North American Benthological an exotic cray®sh: The roles of competition and predation. Ecolo- Society 18, 74±98. gical Applications 9, 678±690. Stylinski, C.D., Allen, E.B., 1999. Lack of native species recovery fol- Holway, D.A., 1999. Competitive mechanisms underlying the dis- lowing severe exotic disturbance in southern California shrublands. placement of native ants by the invasive argentine ant. Ecology 80, Journal of Applied Ecology 36, 1±12. 238±251. Suarez, A.V., Bolger, D.T., Case, T.J., 1998. E€ects of fragmentation Ingold, D.J., 1996. Delayed nesting decreases reproductive success in and invasion on native ant communities in coastal southern Cali- Northern Flickers: implications for competition with European fornia. Ecology 79, 2041±2056. Starlings. Journal of Field Ornithology 67, 321±326. Toft, R.J., Rees, J.S., 1998. Reducing predation of orb-web spiders by Kaiser, J., 1999. Stemming the tide of invading species. Science 285, controlling common wasps (Vespula vulgaris) in a New Zealand 1836±1841. beech forest. Ecological Entomology 23, 90±95. J.C. Burger et al. / Biological Conservation 98 (2001) 241±249 249

Uetz, G.W., Unzicker, J.D., 1976. Pitfall trapping in ecological studies Wise, D.H., 1993. Spiders in Ecological Webs. Cambridge University of wandering spiders. Journal of Arachnology 3, 101±111. Press, New York. Westman, W.E., 1979. Oxidant e€ects on Californian coastal sage Zedler, P.H., Gautler, C.R., McMaster, G.S., 1983. Vegetation change scrub. Science 205, 1001±1003. in response to extreme events: the e€ects of a short time interval Westman, W.E., 1981. Factors in¯uencing the distribution of species between ®res in California chaparral and coastal sage scrub. Ecol- of Californian coastal sage scrub. Ecology 62, 439±455. ogy 84, 809±818.